This joint webinar with Databricks introduces AIStor Table Sharing, a native Delta Sharing implementation embedded directly at the storage layer with no sidecar services required. The session makes the case that hybrid data environments don't need replication pipelines: Databricks can query live on-premises data in AIStor through a read-only, token-authenticated share that surfaces in Unity Catalog like any native table. A live demo covers share creation in the AIStor console, token generation, Unity Catalog import, live SQL joins across shared Iceberg tables, and dynamic table addition without reissuing credentials. Dennis Dau from Databricks covers the Delta Sharing protocol, its open standard foundation, and how AIStor Table Sharing enables access to the Databricks marketplace data collaboration platform and clean room ecosystem for on-premises data estates.
.svg)
A customer benchmark showed the same query completing in 4 seconds via AIStor Table Sharing versus over 40 seconds through their ETL and replication process, a 10x improvement in time-to-result, over a 2 Gbit WAN connection with 15ms latency.
AIStor is the first storage platform to embed Delta Sharing natively at the storage layer; MinIO was the first on-premises implementation and Databricks noted that joint customers are already using it in production today.
AIStor Table Sharing unlocks on-premises data for the Databricks marketplace data collaboration platform and clean room ecosystem, including regulated industries like financial services where Mastercard is cited as an active customer.
Data and analytics leaders, platform architects, and IT decision-makers at enterprises with significant on-premises data estates who need to give Databricks cloud analytics access to that data without replication, ETL pipelines, or data sovereignty trade-offs.
.svg)
.jpg)
